Thermal Desorption Technical Support

Thermal Desorption Technical Support Note 47: The Analysis of Landfill Gas Compounds using Thermal Desorption GC/MS and a Retention Time Locked Database Key Words: VOCs, directive, compliance, identifi cation Introduction As a consequence of their content and decomposition processes, landfill sites containing domestic and commercial waste produce a variety of odorous and toxic volatile organic compounds (VOCs) by a time dependent process. A recent European Directive - 1999/31/EC - on Landfill of Waste, specifies the following actions to be taken by Local Authorities involved in the disposal of waste to landfill sites. Under the directive: Landfill sites must be classified as either hazardous, non-hazardous (municipal) or inert Monitoring requirements for all sites are specified as follows: The volume and composition of leachate The composition of both ground water and the water of any nearby streams (if applicable) Monitoring of bulk gases (e.g. CO 2, CH 4, O 2 ) should be carried out monthly while the landfill site is in operation and 6-monthly during the after-care phase. Other gases must be monitored as required (see below). Emissions of odours and dust must be minimised. All except inert landfill sites must be monitored for the priority pollutants (see Table 1) at least annually or more frequently if required. Most of the priority pollutants are volatile organic compounds (VOCs). In the UK, a standard sampling and analytical protocol for compliance with the monitoring requirements of this directive is currently being finalised by the Environment Agency. It specifies pumped sampling onto 2-bed Silcosteel sorbent tubes followed by TD-GC/MS analysis. Other European countries are also in the process of drafting compliance protocols. Objectives As well as the 20 target compounds listed in Table 1, landfill gas may contain several hundred VOCs and other gases. It was therefore necessary to develop a complete analysis method (thermal desorption - gas chromatography - mass spectrometry (TD GC/ MS)) that would be able to sample, analyse and positively identify the presence (or absence) of each of these target compounds. Using the unique features available within the Markes ULTRA-UNITY TD system linked to the Agilent 6890 GC 5973 MS with Chemstation software the following objectives were met: Develop a conventional tube sampling protocol for the landfill gas Develop a TD method for analysis of the 20 target compounds Develop a GC/MS analytical method for maximum compound separation in time by mass Create a Retention Time Locked (RTL) method Produce an RT Locked target compound database

Page 2 of 10 1) Vinyl Chloride 2) 1,3-Butadiene 3) Methyl Mercaptan 4) Chloroethane 5) 1-Pentene 6) Furan 7) Ethyl Mercaptan 8) 1,1- Dichloroethene 9) Dimethylsulphide 10) Carbon Disulphide Table 1: Priority Toxic & Odorous Compounds Method Sampling 11) 1,2-Dichloroethene 12) 1,1-Dichloroethane 13) Propyl Mercaptan 14) Tetrachloromethane 15) Benzene 16) Trichloroethene 17) Butyl Mercaptan 18) Dimethyldisulphide 19) Ethylbutyrate 20) 2-Butoxyethanol The nature of landfill gas poses a number of problems when sampling onto sorbent tubes for thermal desorption GC (/MS) analysis. Firstly, the labile nature of many of the target compounds precludes the use of ordinary stainless steel sorbent tubes, as many of the compounds of interest would break down when heated in these tubes. Therefore tubes which have been specially treated with a silica based inert coating must be used. These tubes have an extremely thin (Angstroms) inert coating bonded onto the stainless steel surface of the tube. The tube, and the sorbent retaining gauzes used at the front of the tube and between the different sorbents are all coated. These inert tubes have been shown to cause little or no decomposition of labile compounds during the sampling and thermal desorption processes (further information can be found in TDTS 14). The target compound list also presents challenges with respect to the choice of sorbent material used to trap the compounds. The sorbent(s) used must be fully inert, in order not to compromise the thermally labile compounds, and must be suitable for a wide volatility range such as is found in the priority list. The sorbent mixture of choice is therefore Tenax TA which is suitable for trapping the less volatile compounds such as benzene, backed up by a bed of UniCarb (formerly known as Spherocarb) which is a strong, carbonised molecular sieve sorbent, suitable for trapping the volatile compounds such as vinyl chloride. Both these sorbents are inert and have very low background artifact levels (<1.0 ng Tenax, <0.1 ng UniCarb). In addition landfill gas has a very high humidity and is usually at an elevated temperature, which leads to relatively high levels of water being trapped on the tubes and therefore to the possibility of sample breakthrough and loss of some of the more volatile compounds. In order to reduce the amount of water trapped on the tube sample volumes should be restricted to 100 ml. To further remove excess water, each tube must be dry purged. Dry purging the sample tube simply involves passing a volume (in this case typically ~400 ml) of pure, dry, air or inert gas through the tube from the sampling end at a rate of ~50 ml/min, prior to analysis. Care must be taken not to exceed the breakthrough volumes for any of the retained analytes during the dry purge process. Thermal Desorption Method A 3-stage thermal desorption method was developed (details are shown in Figure 1). A two-stage tube desorption method was used in order to desorb the labile sulphur compounds at the lower temperature before desorbing the remaining compounds at 300 C - a temperature which may have caused some degradation of the labile compounds. The cold trap was also packed with Tenax TA and Unicarb. The trap low temperature was set to 30 C which allowed most of the water present in the sample to purge through the trap. A trap heating rate of 40 C/s was used (rather than the 100 C/s default rate) to give a more controlled release of compounds from the trap, again to ensure that there was no degradation of the labile compounds.

Page 3 of 10 two stage tube desorption parameters electronic split flow control trap desorption values Figure 1: Thermal Desorption Method GC/MS Method The column used was a 60 m, 0.25 mm I.D. 1.4 µm DBVRX column which was selected for optimum resolution of volatiles. GC/MS conditions were as follows: Carrier Gas: Helium GC Temp 1: 40 C Ramp Rate: 10 C/min GC Temp 2: 225 C MS Mode: Scan Mass range: 35-260 amu Threshold: 50 Scans/Sec: 3.25 Injection of Standard Mix A 50 ppm liquid standard containing the 20 target compounds was used to develop the chromatographic and MS method for optimal compound separation (Figure 2). Where complete, chromatographic resolution was not possible, compound identity was achieved from extracted ion profiles. Future quantitation would ultimately use this as well (Figure 3). All the priority toxic and odorous compounds listed in Table 1 were identified, including the labile thiols and sulphides (Figure 4). The 50 ppm standard was injected as a liquid, using the calibration solution loading rig (CSLR), onto a TD tube to be loaded into the ULTRA- UNITY system. However, real samples from the landfill site are taken as gas samples. To check that this would not affect the retention times of the target compounds an injection was made from the headspace above the methanol standard and the retention times of the target compounds was found to be the same (Figure 5).

Page 4 of 10 Vinyl Chloride Methanethiol Ethyl Chloride Ethanethiol 1-Pentene Furan 1,1 Dichloroethene / DMS Carbon Disulphide 1,2 Dichloroethene (t) 1,1 Dichloroethane 1,2 Dichloroethene (c) 1-Propanethiol 1,1,1-Trichloroethane Carbon Tetrachloride Benzene 1-Butanethiol Dimethyldisulphide 1,1,2 Trichloroethane Ethyl butyrate 2-Butoxyethanol Figure 2: A 1 µl injection of a 50 ppm standard mix Total Ion Chromatogram Extracted ion profiles 38 amu - Furan 62 amu - Ethanethiol Figure 3: Total Ion Chromatogram and Extracted ion profiles of furan and ethanethiol

Page 5 of 10 4.77, 4.79 mins 8.41, 8.41 mins 16.23, 16.24 mins 1 µl methanol standard 300 ml headspace from methanol standard vinyl chloride dimethyl sulphide benzene Ethanethiol Methanethiol Propanethiol Butanethiol Dimethyldisulphide DMS Figure 4: Extracted ion profile at 47 amu - thiols and sulphides Figure 5: Comparison of standard by liquid injection and standard by headspace injection

Page 6 of 10-20% P -10% P column pressure = P +10% P +20% P Target Peak (1,1-Dichloroethane) Spectra Figure 6: 5 data files acquired at different pressures Compound Identification A conventional MS (PBM) database was created for each target compound from the standard solution. With very complex chromatograms, such as those obtained from real landfill gas samples, target peaks can be obscured by other coeluting, or partially co-eluting compounds. Positive identification in such complex samples can be further complicated by tiny shifts in the retention times of target compounds due to small changes in carrier gas pressure each time the GC system is set up. Therefore it is necessary to look for further levels of compound identification in order to confirm the presence or absence of a particular target analyte. To this end, a retention time locked database can be created. Using the 50 ppm standard as previously, five data files were acquired at different carrier gas pressures (±20%, ±10% and method value) (Figure 6) and retention time was calibrated against pressure. A target peak was specified - in this case 1,1 dichloroethane, and the retention time defined as given in the Retention Time Locking Calibration Report (Figure 7). The method is then locked to the target peak retention time. The landfill MS library created from the standard solution is then converted to an RTL Target Database (Table 2) where against each target compound is listed a target ion, an expected retention time and 3 qualifying ions - these ions can be edited by the user if required. The RTL Database is then assigned to a method. Analysis of real landfill samples 100 ml of Landfill gas sample was drawn through a sorbent tube packed with Tenax TA and UniCarb, and the tube analysed using the thermal desorption method developed above. The chromatogram is shown in Figure 8. The Chemstation retention time locked method then produced a Retention Time Locked Screen Report detailing which target compounds had been found shown in Table 3. The report for this analysis is shown below. It consists of an identification status, X for a definite hit.? for a partial hit; the expected retention time and deviation DRT; the target ion mass and response; qualifiers out of range which results in a partial (?) indentification; and finally

Page 7 of 10 Retention Time Locking Calibration Report Retention Locked Method: C:\MSDCHEM\1\METHODS\LANDFILL.M Retention Locked Cal Date: 26 Mar 2003 4:48 pm Instrument: 6890 GC/MS Maximum Deviation: 66.894 seconds Operator: Gareth Roberts RTL Curve: R = 1.66e+004 A*A - 6.77e+005 A + 7.14e+006 Method Lock is currently On Compound: 1,1-Dichloroethane Retention Time Calibration: File psi pressure Time min Spec Xcor Deviation Seconds RTLOCK1.D 17.91 12.611 0.99974 66.894 RTLOCK2.D 20.15 12.032 0.99777 32.178 RTLOCK3.D 22.39 11.496 1.00000 0.000 RTLOCK4.D 24.63 10.995 0.99925-30.054 RTLOCK5.D 26.87 10.507 0.99830-59.310 Terms of Curve Fit: Constant = 7.14021e+006 Linear = -677432 Quadratic = 16583 Coefficient = 0.999986 ** Good Fit ** Locked Retention Time information: Retention Locked File: C:\MSDCHEM\1\METHODS\LANDFILL.M\rtlock\rtlock3.d Acq Date: 26 Mar 2003 4:48 pm Instrument: 6890 GCMS Operator: Gareth Roberts Measured Retention Time: 11.496 Pressure: 22.39 (psi) Locked Retention Time: 11.496 Pressure: 22.39 (psi) ** Locked RT and Pressure Within Calibrated Limits ** Lock run spectrum XCor: 1.0000 Report created: Mon Apr 07 13:51:07 2003 Figure 7: Retention Time Locking Calibration Report the XCR value which is a number the library. A low XCR value should not be taken as nonidentification, it could be the correct compound but occluded by a co-eluting non-target compound, producing a poor spectral match. From this report it is evident that 8 compounds were definitely found and 5 were absent. With the remaining 10 compounds it was not possible to state with complete certainty whether they were present or absent from the sample. This is almost certainly due to the complexity of the real sample compared with the standard. Further confirmatory compound identification using spectral deconvolution (AMDIS) and NIST searching software can be used to obtain additional positive identification (see TDTS 62). Conclusions A complete thermal desorption GC/MS method has been successfully developed for the highly complex mixture of toxic and priority odourous VOCs found in landfill gas. Markes International inert coated 2-bed sorbent tubes were used to take small (100 ml) samples of landfill gas and the tubes were dry purged prior to analysis to remove excess water. Using the power of the Agilent GC/MS Chemstation, the presence (or absence) of target compounds was confirmed using a Retention Time Locked screenable database. This multiple level of identification confirmation is necessary when analysing such complex mixtures as those found in landfill gas sites. Additional confirmatory techniques can be used for identification which utilises different library search algorithms, such as AMDIS, NIST etc.

Page 8 of 10 Cpd# Compound Name TIon Exp_RT Q1 Q2 Q3 1 Ethene, chloro- 62 4.79 27 64 26 2 1,3-Butadiene 39 5.07 54 53 27 3 Methanethiol 47 5.32 48 45 46 4 Ethyl Chloride 64 5.85 66 49 26 5 1-Pentene 42 7.10 55 41 70 6 Furan 98 7.58 39 38 40 7 Ethanethiol 82 7.64 29 47 34 8 Ethene, 1,1-dichloro 91 8.36 96 98 63 9 Dimethyl sulfide 92 8.42 47 45 46 10 Carbon disulfide 76 9.26 44 78 77 11 Ethene, 1,2-dichloro-, (E)- 61 10.80 96 98 26 12 Ethane, 1,1-dichloro- 63 11.51 65 27 83 13 Ethene, 1,2-dichloro-, (Z)- 61 13.05 96 98 63 14 1-Propanethiol 76 13.20 43 47 42 15 Ethane, 1,1,1-trichloro- 97 15.33 99 61 117 16 Carbon Tetrachloride 117 16.13 119 121 82 17 Benzene 78 16.24 77 51 50 18 Trichloroethylene 95 17.63 130 132 97 19 1-Butanethiol 41 18.10 56 27 90 20 Diusulfide, dimethyl 94 19.51 79 45 46 21 Ethane, 1,1,2-trichloro- 97 20.18 83 99 85 2285 Butanoic acid, ethyl ester 71 21.14 43 88 27 2327 Ethanol, 2-butoxy- 57 24.18 45 87 56 Table 2: Landfill MS Library converted to RTL Target Database

Page 9 of 10 Figure 8: 100 ml landfill gas sample

Page 10 of 10 Compound Status ExpRT DeltaRT Tgt Tgt Response Q ions out of Range 1 Ethene, chloro-? 4.79 0.008 62 1790570 27,26 0.6 2 1,3-Butadiene? 5.07-0.099 39 2684920 7 54, 53, 27 0.06 3 Methanethiol 5.32 47 -Not Found- 4 Ethyl Chloride? 5.85-0.041 64 118661 6 26 0.64 5 1-Pentene? 7.10 0.046 42 301883 0 55, 41, 70 0.06 6 Furan x 7.58-0.066 98 289780 0.97 7 Ethanethiol 7.64 82 -Not Found- 8 Ethene, 1,1-dichloro-? 7.86-0.050 91 147586 96, 98 0.39 9 Dimethyl sulfide x 8.42-0.094 92 209966 0.82 10 Carbon disulfide x 9.26-0.078 76 266235 0.98 11 Ethene, 1,2-dichloro-, (E)? 10.8-0.099 61 160590 26 0.49 12 Ethane, 1,1-dichloro-? 11.51-0.102 63 299526 27 0.97 13 Ethene, 1,2-dichloro- (Z) x 13.05-0.099 61 3006044 0.99 14 1-Propanethiol 13.20 76 -Not Found- 15 Ethane, 1,1,1-trichloro x 15.33-0.092 97 79064 0.99 16 Carbon Tetrachloride 16.13 117 -Not Found- 17 Benzene x 16.24-0.087 78 1570194 0.99 18 Trichloroethylene x 17.63-0.099 95 769325 0.99 19 1-Butanethiol? 18.10-0.021 41 210295 56, 27, 90 0.15 20 Disulfide, dimethyl x 19.51-0.099 94 215921 0.98 21 Ethane, 1,1,2-trichloro- 20.18 97 -Not Found- 22 Butanoic acid, ethyl ester? 21.14-0.093 71 6419284 27 0.88 23 Ethanol, 2-butoxy-? 24.18-0.131 57 548700 45, 46 0.79 Table 3: Retention Time Locked Screen Report XCR Applications were performed using the stated analytical conditions. Operation under different conditions, or with incompatible sample matrices, may impact the performance shown. For More Information These data represent typical results. For more information on our products and services, visit our Web site at www.agilent.com/chem. www.agilent.com/chem Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Agilent Technologies Published July 19, 2013 5991-2825EN Markes International Ltd. T: +44 (0) 1443 230935 F: +44 (0) 1443 231531 E: enquiries@markes.com